Anthropic's Fable 5.2 has demonstrated a significant leap in completing real-world freelance tasks, achieving 41.3% success rate at professional quality. This performance more than doubles that of its closest competitors, Opus 5.5 and GPT 6 Astra. A year prior, the leading models could only complete 2.5% of such jobs, indicating rapid progress in AI's capability to handle complex, real-world assignments, though areas like product design and animation still present challenges. AI
IMPACT Sets a new benchmark for AI in completing real-world freelance tasks, signaling accelerated adoption in professional services.
RANK_REASON Frontier-lab model release with system card [lever_c_demoted from frontier_release: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
